For the first year, a blog is a list of things you have written and the only question is what to write next. Somewhere around the twentieth post that changes. You now have a library, some of it is competing with itself, some of it is out of date, and the highest-value work is no longer adding to it.
Very few small business blogs ever make that switch. They keep publishing into a pile nobody maintains, and the pile gradually stops working.
The short answer
Once a year, sort every post into four groups: working (leave alone), nearly working (improve these — biggest return available), competing with each other (merge into one), and never going to work (delete or redirect). An afternoon on this typically beats an afternoon of new writing.
Posts that compete with each other
The most common problem in a library that grew organically, and the least obvious.
Over two years you write "what does a new fence cost", then "fence pricing explained", then "how much should I budget for fencing". Three posts, three different months, all answering the same question. It feels like thorough coverage. It is not.
Google has to choose one of them, and it frequently chooses badly or alternates between them. Worse, whatever credibility each page earns is split three ways instead of concentrated in one, so all three sit lower than a single strong page would.
The fix is merging. Pick whichever has the best search history, move the genuinely useful parts of the other two into it, then redirect the old URLs to the survivor so anything they had earned flows into it. One page covering the question properly outranks three covering it partially.
To spot these, open Search Console, look at the queries bringing you impressions, and check whether two of your pages show up for the same one. That is the signal.
Posts that are nearly working
Every mature blog has several pages averaging around position eleven to fifteen. Shown regularly, clicked rarely, sitting just below where anybody looks.
These are worth more than new posts, by a wide margin. Getting a page from twelfth to sixth multiplies its traffic several times over, and it is far less work than taking a new page from nothing to sixth.
What usually moves them: answering the question more directly in the first paragraph, adding the related sub-questions people also search, putting in specifics that the pages above you lack — real numbers, real local detail — and linking to the page from more of your own posts.
One of these a quarter is a genuinely good use of time, and almost nobody does it, because editing an old post does not feel like progress in the way a new post does.
Posts that are never going to work
Harder, because you wrote them and they are fine.
A post with two years of history, no impressions worth mentioning and no clicks is not waiting for its moment. It is targeting something nobody searches, or competing with national operators, and more patience will not change either.
Options, in order of preference. If it contains anything useful, fold that into a related post that is working and redirect the URL there. If it does not, and it is harmless, leave it — a few dead pages hurt nothing. Delete outright only if it is genuinely embarrassing or actively wrong, and redirect the URL to the nearest relevant page rather than leaving a dead end.
Resist large-scale pruning. Advice about "content pruning" was written for sites with thousands of pages; on a blog of thirty it mostly just removes things.
Posts that are working
Leave them alone, with one exception.
There is a real temptation to keep improving your best page. Usually this makes it worse — additions accumulate, the direct answer at the top gets buried under context, and the thing that made it work erodes.
The exception is factual decay. Prices from two years ago are wrong now. So are timeframes, regulations, and any sentence containing "recently". Update those, leave the structure alone, and change the date only when you have genuinely changed something.
The one structural thing worth doing
Beyond the four groups: link your posts to each other properly, once.
Most libraries are a flat list where every post links out to a service page and none link to each other. Adding two or three internal links per post — to the genuinely related ones, in the body where they make sense — has a disproportionate effect. Newer pages get found faster, older pages get reinforced, and readers stay for a second article instead of leaving after one.
An afternoon on a twenty-post blog. It is frequently the single biggest improvement available, and it costs nothing but the afternoon.
Keep the library pointed somewhere
As the collection grows, the original discipline tends to slip: posts start getting written because a topic came up rather than because a page needed support.
Check that every post still links to somewhere a reader could hire you, and that the destinations still reflect what you want more of. Businesses change — the service you were pushing two years ago may not be the one with capacity now. Re-pointing a cluster of posts at a different page is a ten-minute job with a real effect, and it is the sort of thing that only makes sense once you think of the blog as a set of pages supporting destinations rather than a chronological feed.
An afternoon a year
None of this is difficult and none of it is urgent, which is exactly why it never happens.
Put it in the calendar annually, in a quiet month. Open Search Console, sort the pages by impressions, and work down the list deciding which of the four groups each one belongs to. Two hours, maybe three.
The result is nearly always a better return than the same hours spent writing something new, because a library that is maintained keeps compounding and one that is only added to gradually stops.